troubleshooting Input interface errors on VL

[rikdrt1]

i have a 6807 core switch setup as a VSL-Pair.   have about 1000 VoIP phones throughout maybe 10-12 access switches.   in the last few weeks i have been noticing that people on voice-VL1 having issues with dropped calls and the other folks on voice-VL2 are fine.  i went to check a few common items first then as i poked around further i ran some basic SolarWinds utilities to see why this was happening and noticed 5 devices on the problem VL were intermittently unavailable and getting packet loss from my simple utility.  the other 5 on the other VL was not having any issues and i can ping all 5 of those devices with no errors.

so i checked the VL interfaces and found this..... (below)

wondering how this can happen only on 1 of the 30 VL's on this switch.     bad phone/device causing errors maybe ?

any help would be appreciated.  thanks.

--------------------------

b8m-s01#  sh int vlan 108 | i drops
Input queue: 0/75/123326/0 (size/max/drops/flushes); Total output drops: 0
0 unknown protocol drops
b8m-s01#sh int vlan 92 | i drops
Input queue: 0/75/0/0 (size/max/drops/flushes); Total output drops: 0
0 unknown protocol drops

[MHM Cisco World]

Your QoS drop tail is config only for one vlan other is not so that why the QoS drop packet before it full the queue.

Share both qos of both vlan 

[balaji.bandi]

First, what is the difference between VL1 and VL2 - compare? is this the same issue on the same Stack or a different stack?

what is the Layer 3 Gateway? for this VLAN on Core Switch 6807 ?

Coming back to the Physical Location, is this issue device all in one location or 1 switch?

Do all Access switches have the same config?

First, i start with the Fresh Patch cable, Same switch try different ports for 1 device and test it.

I compare the configuration.

by the way what model of access switches?

[rikdrt1]

Same 6807-XL VSL pair.

Just two different VLANs on the same device.

I did test and replace all the later1 items, including the new 7841 voip phones.

But when I moved it to the clean VL92, it works flawlessly. Put it on the other one, and I see errors and disconnects for all 300+ phones on that VL. So the config is the same - 3 lines on the VL config. Just the IP and the helper-address and that's it.

[balaji.bandi]

are the end devices connected to 6807 ? or what is the access switch? what is the Access switch model ?

VL1 means native VLAN 1 ?

[rikdrt1]

Yes VLAN 92 and VL108
All the access switches connect to the 6807’s most are 4500 class and some of the newer 9300/9400’s. A total of about 20+. As far as the config, they are or suppose to be the same but that’s a good point I should check to make sure .. especially in the newer area where this new VL108 is primarily used.
Some are 1G uplinks in a port channel config and others as 10G

All of the switch ports have switchport voice vlan 108 setup Or switchport voice vlan92 which is clean

[balaji.bandi]

b8m-s01#  sh int vlan 108 | i drops
Input queue: 0/75/123326/0 (size/max/drops/flushes); Total output drops: 0
0 unknown protocol drops
b8m-s01#sh int vlan 92 | i drops
Input queue: 0/75/0/0 (size/max/drops/flushes); Total output drops: 0

what device did you take the output from Cat 4500 or Cat 9K ?

[rikdrt1]

This came from the core switch 6807XL where all the VL are defined. I didn’t check each access switch yet but I will in the coming days. But since I saw it on the core I assumed it’s the source ? Or is that not the case ?

[balaji.bandi]

At this, it is hard to say what is the issue, you are just relying on only Ping- as I understand your setup VSS holds all the SVI, Stack of Access switches only Layer 2.

If this only certain Stack, i will be more focused to Look at those switches compare to others working as expected.

This case Model, IOS Code? what is the speed of the interface, what if the device has an issue plugged into the working switch, is these expected results the same?

[rikdrt1]

Thanks. I haven't dug that deep yet , only because I was thinking this was a core switch issue, but basically what I did test so far is this:

VL92(clean counters) exists on about 10 access switches, let's call that the original older area... as the building expanded we added more user/PC VLANs and in the newer area there is another /23 network - VL108(with the errors) with about 10-12 access switches of all kinds (3850/9200/9300/4510).



Several folks were on a 4510R/VL108 , and were experiencing dropped calls and solarwinds was reporting packet loss to the 2-3 test phones in that newer area. After I changed cables, and even phone devices - no changes. So I went to this access switch and moved the voice VL from 108 to 92 and issue was resolved. I still see the remaining VL108 phones on that same swtich intermittently have some packet loss throughout the day but the two phones I moved to the clean VL are perfect at 0% packet loss all day long.

[MHM Cisco World]

can you share

show queueing interface

[rikdrt1]

sh queueing int vlan 108

Interface Vlan108 queueing strategy: none